AM Intelligence Orders 9,000 NVIDIA Rubin GPUs to Launch Asia’s First Frontier AI Cluster in Hyderabad

AMI has placed binding order for ~9,000 NVIDIA Rubin GPUs deployed as Vera Rubin NVL72 rack-scale systems in its first AI Factory in Hyderabad, part of its 1 GW compute capacity globally

AM Intelligence ("AMI"), the AI infrastructure platform set up by Promoters of Greenko, is striving to become the World’s leading, intelligence infrastructure company, powering the AI economy through vertically integrated energy, compute and innovation.

AMI today announced first step towards this vision by placing a firm and binding order for Vera Rubin delivery in Q1 2027 at its first AI factory in Hyderabad. The order covers 9,000 NVIDIA Rubin GPUs deployed as Vera Rubin NVL72 rack-scale systems for the 30 MW of capacity and will make AMI’s Hyderabad AI Factory as one of the first frontier AI compute clusters in Asia.

The Hyderabad facility will be the first tranche of AMI's planned 1 GW of Compute-as-a-Service capacity across India, the United States, Finland and Malaysia. AMI plans to bring an initial 200 megawatts (MW) of capacity to market in the near term, involving capital expenditure of more than USD 8 billion.

Anil Chalamalasetty, Chairman, AM Intelligence, said: “For over two decades, we have focused on transforming electrons into value — addressing critical and emerging societal needs. Today, the electron-to-token opportunity allows us to take this capability further, converting power infrastructure into frontier AI compute at scale. Bringing latest generation NVIDIA Vera Rubin to India marks the next step in this journey and establishes AMI at the forefront of the emerging AI electron-to-token economy.”

By integrating NVIDIA Vera Rubin architecture with high-throughput RDMA over Converged Ethernet (RoCE) networking for fast, efficient data movement, and advanced storage systems, the Hyderabad AI Factory is being designed to run some of the world’s largest and most sophisticated AI models, including trillion-parameter models and next-generation agentic AI applications.

NVIDIA Vera Rubin NVL72 introduces NVFP4, a new low-precision computing format designed to run AI workloads more efficiently, alongside next-generation High Bandwidth Memory (HBM4). The architecture is designed to reduce AI inference token-serving costs by up to 10 times compared with the preceding NVIDIA Grace Blackwell generation. AMI’s Hyderabad AI Factory is being engineered to deliver approximately 450 exaFLOPS of NVFP4 inference compute, providing massive computing capacity for running advanced AI models at scale.

AMI: Building Toward 5 Gigawatts

Over the last two decades, our focus has been on electron valorisation, creating progressively greater value from energy across energy management, molecules and metals, and now AI tokens, the units of information processed and generated by AI models. This end-to-end approach has been central to our philosophy of addressing critical and emerging societal needs.

Today, intelligence has reached the status of permeating humankind’s lives on a daily basis. AMI, as a strategic builder of long-term heavy assets, will leverage its unique power solutions, capex advantage and access to next-generation NVIDIA AI infrastructure to deliver highly cost-competitive electron-to-token economics by integrating energy infrastructure with the computing capacity required to train and run AI models. This will enable wider deployment of AI with a collaborative service offering spanning across hyperscalers, neo cloud providers, sovereign AI initiatives, frontier labs, AI natives and local developer communities.

AMI is developing 5 GW of powered AI data centres across India, the United States and Europe. These facilities are being designed for high-density AI computing, using liquid cooling to manage the significant power and heat requirements of advanced AI chips, while allowing the infrastructure to accommodate successive generations of AI silicon. AMI also intends to bring 1 GW of Compute-as-a-Service capacity to global AI workloads with initial capacity of 200 MW coming online in short-term.

About AM Intelligence:

AM Intelligence (AMI) is building a full-stack AI infrastructure ecosystem spanning energy, data-center infrastructure, hardware, and customized AI models. Backed by Greenko's owned renewable generation and storage, AMI is developing gigawatt-scale, liquid-cooled compute capacity designed to serve global hyperscalers, frontier AI labs, enterprises, and India's sovereign AI initiatives — while widening chipset access for Indian developers building domestic and global AI solutions. AM Intelligence is developing 5 GW of powered AI data centres and 1 GW of Compute-as-a-Service across India, United States and Europe, with 200 MW of Compute-as-a-Service capacity coming to market in the near term.

AMI, promoted by the founders of Greenko, is building globally differentiated platforms across energy, molecules and AI with a focus on enabling next generation of intelligent infrastructure. On electrons, we are building a 50 GW energy infrastructure platform across 20+ states, spanning solar, wind, hydro and energy storage. We are also developing the world’s first interconnected energy storage platform, targeting 200 GWh of storage capacity across India.

On molecules, we are developing low-carbon molecule platform, including green ammonia, with a target of 5 Mtpa of green ammonia capacity across multiple locations in India. Our 1 Mtpa green ammonia project currently under construction is expected to be among the world’s largest RFNBO-compliant green ammonia facilities, supporting energy diversification for Global markets for multitude of use cases across power generation, mobility amongst others.

Intel Core Ultra Series 3 Arrives: First 2nm ‘Panther Lake’ Chips Push AI PCs Forward

At CES 2026, Intel unveiled the Core Ultra Series 3, its first processors built on the Intel 18A process node. These chips, codenamed Panther Lake, mark a major milestone in Intel’s roadmap, introducing RibbonFET gate-all-around transistors and PowerVia backside power delivery to achieve significant gains in efficiency and performance. The Series 3 lineup is positioned as Intel’s flagship AI PC platform, with over 200 designs expected worldwide, and represents the company’s most ambitious step into AI-native computing.

For an uninitiated, Intel’s 18A process node is Intel’s most advanced semiconductor manufacturing technology, a 2nm‑class node that introduces two major innovations: RibbonFET (gate‑all‑around transistors) and PowerVia (backside power delivery). Together, they deliver higher performance per watt, greater transistor density, and improved scalability for AI and high‑performance computing.

An Intel® Core™ Ultra series 3 processor. (Credit: Intel Corporation)

The processors deliver up to sixteen cores, including twelve efficiency cores, and show dramatic improvements compared to the prior Lunar Lake generation. Intel claims a sixty percent boost in multithreaded performance, seventy-seven percent faster gaming, and battery life extending up to twenty-seven hours in laptops. AI acceleration is central to the design, with up to eighty total TOPS of compute, fifty NPU TOPS, and one hundred twenty GPU TOPS through integrated Xe3 graphics. Memory support reaches ninety-five gigabytes, while maintaining full x86 compatibility for existing applications. The new Core Ultra X9 and X7 headline the range, targeting both creative professionals and gamers.

Compared to Lunar Lake, which was built on Intel 20A and focused on AI-centric V-series designs, Panther Lake represents a broader leap. It combines higher core counts, stronger GPU performance, and longer battery life, while shifting Intel’s competitive positioning against rivals like AMD and Apple, who are also advancing their own AI-focused architectures. Intel emphasized that these chips are designed and manufactured in the United States, underscoring its strategy of semiconductor independence and leadership in advanced nodes.

In essence, Core Ultra Series 3 is not just another generational upgrade—it signals Intel’s transition into the AI-first PC era, blending cutting-edge transistor engineering with practical gains in performance, efficiency, and ecosystem adoption.

Mass Production of World's First Non-Binary AI Chip Marks a New Era in Computing

China has commenced mass production of the world’s first non-binary AI chip, a groundbreaking development that challenges traditional computing limitations. Developed by Professor Li Hongge’s team at Beihang University, this innovation integrates binary logic with stochastic computing, paving the way for energy-efficient, high-performance AI hardware.

What Is a Non-Binary Chip?

For decades, computers have operated on binary logic, where every calculation relies on sequences of 0s and 1s. While highly efficient, binary computing faces growing challenges in power consumption and adaptability. A non-binary chip introduces Hybrid Stochastic Numbers (HSN) —a fusion of traditional binary numbers with probability-based values. This means that, instead of solely relying on rigid binary operations, these chips leverage randomness to optimize calculations, enhancing efficiency and fault tolerance.

A Solution to Major Tech Roadblocks

This non-binary chip addresses two critical hurdles in computing:

• The Power Wall: Traditional chips consume excessive energy, limiting scalability. Non-binary chips significantly reduce power consumption while maintaining speed.
• The Architecture Wall: Many experimental non-silicon chips struggle to integrate with existing systems. This new technology seamlessly aligns with CMOS-based architectures, ensuring compatibility.

Real-World Applications and Strategic Advantages

China is deploying these chips across various industries, including aviation, industrial control systems, and intelligent displays, enabling real-time AI processing with superior efficiency.

Moreover, the chip’s domestic production circumvents U.S. semiconductor export restrictions, reinforcing China’s push for technological self-reliance. The U.S. has imposed strict export restrictions on Nvidia’s AI chips, including the H20 model, which was specifically designed to comply with earlier regulations but is now banned. With China developing its own advanced AI chips, it can bypass these restrictions and continue AI development without relying on U.S. technology.

What’s Next?

This breakthrough could reshape the future of AI hardware, creating faster, smarter, and more energy-efficient systems. As global competition in semiconductor technology intensifies, non-binary computing may soon become the new standard.

Could this revolutionize AI-powered industries? Comment below to have your opinion.... 

IISc Researchers Develop Brain-inspired Computing Platform That Can Store and Process Data

Researchers at the Centre for Nano Science and Engineering (CeNSE) of the Indian Institute of Science (IISc) have developed a groundbreaking brain-inspired analog computing platform. This platform can store and process data in an impressive 16,500 conductance states within a molecular film. This innovation mimics the human brain's neural networks, allowing for more efficient and powerful data processing.

Supported by the Ministry of Electronics and Information Technology (MeitY), the Ministry of Education and the Department of Science and Technology., the team at IISc tapped into tiny molecular movements to design a highly precise and efficient neuromorphic accelerator, which can be seamlessly integrated with silicon circuits to boost their performance and energy efficiency.

Key Features:

High Efficiency: The platform integrates data storage and processing, reducing the need for data transfer and significantly improving energy efficiency.

Advanced AI Capabilities: It can handle complex AI tasks, such as training large language models, on personal devices like laptops and smartphones.

Neuromorphic Design: By using molecular movements to create a "molecular diary," it can access a vast number of memory states, far beyond the binary states of traditional digital computers.

This development could revolutionize AI hardware, making advanced AI tools more accessible and energy-efficient. It's a significant step forward in neuromorphic computing and positions India as a potential leader in global tech innovation.

Published in the journal Nature, this breakthrough represents a huge step forward over traditional digital computers in which data storage and processing are limited to just two states.

Neuromorphic computing differs significantly from traditional computing architectures in several key ways. For an instance, Traditional Computing uses the von Neumann architecture, where the CPU and memory are separate entities. Data is shuttled back and forth between them, which can create bottlenecks. While, Neuromorphic Computing mimics the brain’s neural networks, integrating processing and memory storage in a more interconnected manner, reducing data transfer bottlenecks.

Neuromorphic computing holds great promise for the future, especially in areas requiring high efficiency and adaptability.

Such a platform could potentially bring complex Al tasks, like training LLMs, to personal devices like laptops and smartphones, taking us closer to democratising the development of Al tools.

Neuromorphic computing is a fascinating area. It aims to mimic the neural structure and functioning of the human brain to create more efficient and powerful computing systems. This approach can potentially revolutionize various fields by significantly improving computing efficiency and reducing energy consumption.

Recent advancements in neuromorphic platforms have shown promising results. For instance, these platforms can process information in a way that is more akin to how the human brain works, enabling faster and more efficient data processing. This can be particularly beneficial for applications in artificial intelligence, robotics, and real-time data analysis.

Intel’s Lunar Lake Processors for AI PCs Arriving by September This Year

Intel's Lunar Lake processors, which is the company's upcoming 16th generation CPU and a part of Intel's client processor lineup, are set to make a significant impact in the market with their launch in Q3 2024.

The Core Architecture of Lunar Lake will feature a hybrid core architecture with 4 Performance (P) cores and 4 Efficiency (E) cores, providing a balanced design for both power and performance.

Lunar Lake will power more than 80 new laptop designs across more than 20 original equipment manufacturers (OEMs), delivering AI performance at a global scale for Copilot+ PCs. Lunar Lake will get the Copilot+ experiences, like Recall, via an update when available.

These processors are built on Intel's advanced 20A process node (which is equivalent to 2nm), enhancing performance and power efficiency, especially for mobility devices.

Lunar Lake is expected to be a groundbreaking mobile processor for AI PCs with more than 3 times the AI performance compared with the previous generation. An AI PC has a central processing unit (CPU), a graphic processing unit (GPU) and a neural processing unit (NPU), each with specific AI acceleration capabilities. An NPU is a specialized accelerator that efficiently handles AI and machine learning (ML) tasks right on your PC instead of sending data to be processed in the cloud. The AI PC is increasingly important as the need to automate, streamline and optimize tasks on the PC grows.

Lunar Lake's AI Capabilities is expected to give a significant boost in AI performance, thanks to the inclusion of a new Neural Processing Unit (NPU) capable of over 45 TOPS (Tera Operations per Second), making it four times more powerful than the NPU on Meteor Lake rated at 11 TOPS.

Lunar Lake will also feature the Xe2 GPU architecture, similar to Intel’s upcoming Battlemage discrete GPUs. This includes the Xe Matrix eXtension (XMX) cores, which should improve graphics capability significantly.

Key Features —

Performance: It is claimed to be 40% faster in Stable Diffusion 1.5 running on the GIMP photo editor when compared to the Snapdragon X Elite.

Power Efficiency: It promises the lowest power consumption seen from an x86 processor, which could be a game-changer for battery life in laptops and other portable devices.

Intel's Lunar Lake processors are poised to be a major step forward in the evolution of CPUs, offering improvements in AI, graphics, and power efficiency that could redefine the capabilities of next-generation laptops and other computing devices.

Intel's announcement comes as a competitive response to Qualcomm's Snapdragon X Elite processors, highlighting the intensifying race to dominate the Al and PC market. With such advancements, Lunar Lake processors are poised to power the next generation of Al personal computers, including the Microsoft Copilot+ PCs. 

AI Compute Startup d-Matrix Announces the Opening of India R&D Center at the IESA Vision Summit 2022

Team d-Matrix at IESA Vision Summit 2022 with Mr

India Talent to Engage in Cutting-edge AI R&D

IESA member company, d-Matrix announced at the IESA Vision Summit, the opening of its R&D center in the presence of IESA chairman Vivek Tyagi. Founded by semiconductor veterans from Silicon Valley, Sid Sheth, and Sudeep Bhoja, d-Matrix is building a one-of-a-kind datacentre, AI inferencing platform using a software-first approach coupled with path-breaking hardware innovations in the areas of in-memory computing (IMC) and chiplet level scale-out interconnects. d-Matrix has tackled the physics of memory-compute integration using innovative ML tools, software and algorithms, and circuit techniques, solving the final frontier in AI compute efficiency.

If you are following the evolution of deep learning-powered AI, the renaissance of Generative AI, and the next disruption in computer vision, you likely know it’s all about Transformer based models. They are powering neural nets with billions to trillions of parameters and existing silicon architectures (including the plethora of AI accelerators) are struggling to vary degrees to keep up with exploding model sizes and their performance requirements.

“d-Matrix is addressing the exploding need for more AI compute head on by developing a fully digital in-memory computing accelerator for AI inference that is highly efficient and optimized for the computational patterns in Transformers,” said Sid Sheth, Co-founder, President & CEO, d-Matrix”

At d-Matrix, the team is actively working to build the world’s first inference-focused computing platform for the age of Transformer AI. Transformer-based model architectures are creating a whole new class of models called Generative models that are powering services like language generation, and code generation, d-Matrix has already completed the development of its Nighthawk and Jayhawk platforms that demonstrate the benefits of digital in-memory computing and chiplets for inference compute. The company is now developing the Corsair platform that combines the elements with an open, mature, end-to-end software stack that can be deployed frictionlessly across the cloud to client computing.

“The Corsair chiplet platform provides efficient high bandwidth communication fabric balancing compute, memory & networking to serve generative Transformer AI models like GPT3 & Stable diffusion models at scale,” said Sudeep Bhoja, Co-founder and CTO, d-Matrix.

d-Matrix recently raised Series A funding of $44M from premium investors: Microsoft M12, Playground Global, SK Hynix, and Marvell.

d-Matrix already has R&D centers in Santa Clara and Sydney. When asked about the India center, Sid Sheth said, “We are very excited about our India center. We already have a world-class core team now in design and verification here. The talent in India is just amazing. The India team is going to play a critical role in our growth”

d-Matrix recently roped in Dr. Pradip Thaker to head the India team as VP and Country Head. “We are thrilled to have Pradip join our India team. He brings decades of experience in system-on-chip development” said Sid. Before d-Matrix, Pradip was VP of Engineering and Country Head for the Marvell India team, where he was responsible for a force of 1500+ employees. When asked why he chose d-Matrix, Pradip said, “The opportunity to work on ground-breaking AI technology with the world-class engineering and leadership team at d-Matrix was too exciting to pass up,” said Dr. Thaker. The team has a stellar track record of developing and commercializing silicon systems at scale and has attracted top-tier talent across the industry. We’re going to see AI transform lives in the next decade and d-Matrix is going to play a key part in this revolutionary technology adoption.

Blaize™ Emerges From Stealth to Transform AI Computing; Strengthens Its Presence in India

Blaize, formerly known as Thinci, Unveils the first true Graph-Native silicon architecture and software platform built to process neural networks and enable AI applications with unprecedented efficiency

Expands its capacity in Hyderabad to enable hiring of engineers

• Blaize Graph Streaming Processor^TM (GSP) architecture:
  the first to enable concurrent execution of multiple neural networks and
  entire workflows on a single system, while supporting a diverse range
  of heterogeneous compute intensive workloads
• Fully programmable solution brings new levels of flexibility for
  evolving AI models, workflows, and applications that run efficiently
  where needed, a breakthrough for dynamic intelligence at the edge
• Directly addresses technology and economic barriers to AI adoption
  via streamlined processing that yields 10-100x improvement in systems
  efficiency, lower latency, lower energy, and reduced size and cost

Blaize^TM has emerged from stealth unveiling a ground-breaking next-generation computing architecture that precisely meets the demands and complexity of new computational workloads found in artificial intelligence (AI) applications. In cadence with the development and as part of their growth strategy, the company has expanded its existing capacity in Hyderabad which can now accommodate 450 employees; thus giving more scope for hiring skilled engineers from India.

Driven by advances in energy efficiency, flexibility, and usability, Blaize products enable a range of existing and new AI use cases in the automotive, smart vision, and enterprise computing segments, where the company is engaged with early access customers. These AI systems markets are projected to grow rapidly* as the disrupting influence of AI transforms entire industries and AI functionality becomes a “must-have” requirement for new products.

Blaize was founded on a vision of a better way to compute the workloads of the future by rethinking the fundamental software and processor architecture. It is a special day for the entire team in Hyderabad which has played a crucial role in helping us in realising this vision”, says Dinakar Munagala, Co-founder and CEO, Blaize. “We see demand from customers across markets for new computing solutions that address the immediate unmet needs for technology built for the emerging age of AI, and solutions that overcome the limitations of power, complexity and cost of legacy computing.”

Graph-Native Techniques Drive Huge Efficiency Gains

The scope and capabilities of the comprehensive Blaize technology stack is unprecedented. The Blaize GSP architecture and Blaize Picassoä software development platform deliver breakthroughs in computational efficiency. The solution blends dynamic data flow methods and graph computing models with fully programmable proprietary SOCs. This allows Blaize computing platforms to exploit the native graph structure inherent in neural network workloads all the way through runtime. The massive efficiency multiplier is delivered via a data streaming mechanism, where non-computational data movement is minimized or eliminated. This gives Blaize systems the lowest possible latency, reduces memory requirements and reduces energy demand at the chip, board and system levels.

Blaize GSP is the first fully programmable processor architecture and software platform that is built from the ground up to be 100% graph-native. While there are many types of neural networks, all neural networks are graphs. With the inherent graph-native structure, developers can now build multiple neural networks and entire workflows on a single architecture that is applicable to many markets and use cases. End-to-end applications can be built integrating non-neural network functions such as Image Signal Processing with neural network functions, all represented as graphs that are processed 10-100 times more efficiently than existing solutions. And AI Application developers can now build entire applications faster, optimize these for edge deployment constraints, and run them efficiently using automated data streaming methods.

Quotes from Industry Analysts, Investors and Customers:
“Blaize is very innovative, our important business partner,” says Yukihide Niimi, CEO of NSITEXE and DENSO Advisory Board member. “DENSO is demonstrating leadership in many areas as the automotive industry undergoes extraordinary technology changes. NSITEXE was established to catch up such technology change and to accelerate development of flexible compute IP solutions like DFP. NSITEXE is willing to work together with Blaize to boost the flexible Graph (Data Flow) compute technology ecosystem."

"I have been watching Blaize for several years and saw early on that their graph-native architecture would be particularly well suited to a wide range of AI and robotics workloads,” says Schuyler Cullen, VP AI & Robotics, Samsung Strategy and Innovation Center. “I have been impressed by their rapid scaling in team, organization, and technology."

"The proliferation of AI across multiple industries and application areas is dependent upon robust, programmable, efficient, scalable, high-performance hardware, that extends AI processing from cloud datacenters through to the end device, server or appliance,” says Aditya Kaul, Research Director, Tractica. “It’s becoming clear that traditional processing architectures will not be enough to meet the demands of this new emerging market, with new techniques like graph-based computing showing promise. Success will be defined by combining new computing approaches with modular hardware and a deployment-oriented software stack, all of which is part of the Blaize value proposition from day one."

"Blaize’s vision of a native graph streaming processor (GSP) is relatively unique,” noted Karl Freund, Sr. AI analyst at Moor Insights & Strategy. “The GSP is more general purpose than, say, a single-function ASIC for AI, and can consequently create opportunities in many markets, from Automotive to the Edge to the Cloud."

"The coming out of Blaize and its leading Graph Streaming Processor is extremely exciting,” says David (Dadi) Perlmutter, an angel investor, entrepreneur and former EVP and Chief Product Officer of Intel corporation. “As an initial investor in Blaize, I recognized early on the great efficiency of one of the first to market a complete solution designed from scratch, fully optimized for AI and Neural Network applications. The unprecedented efficiency is great for a wide range of edge applications, particularly the automotive market. I am proud of the team in delivering on the promise."

About Blaize

Blaize leads new-generation computing unleashing the potential of AI to enable leaps in the value technology delivers to improve the way we all work and live. Blaize offers transformative solutions that optimize AI wherever data is collected and processed from the edge to the core, with focus on automotive, smart vision and enterprise computing markets. Blaize has secured US$87M in funding from strategic and venture investors Denso, Daimler, SPARX Group, Magna, Samsung Catalyst Fund, Temasek, GGV Capital, SGInnovate, and Magna. With headquarters in El Dorado Hills (CA), Blaize has teams in Campbell (CA), Cary (NC), and subsidiaries in Hyderabad (India), Leeds and Kings Langley (UK), with 325+ employees worldwide.

-The AI systems market is predicted to reach $97.9 billion by 2023, as AI influence changes entire industries. International Data Corporation (IDC) Worldwide Artificial Intelligence Systems Spending Guide September 2019
-AI functionality will be a requirement for many new products and by 2023 will represent a $34.3 billion revenue opportunity for semiconductor vendors.
Gartner: AI Neural Network Processing Semiconductor Revenue, Worldwide, 2019 Update

~ BusinessWire India